[charlieslasher]

Documenting the adventure of making an eg machine base. Mostly photos and some words.

First test pieces were eg added to 4th axis subplate.

Aggregate dryed for week in the sun then vacume dryed for 4hrs per 40kg in a 60L paint pressure pot.

220kg cast so far
10-12% EPoxy by weight
West 105
West slow hardner
Gerneral builders mix (used for diy concrete) + extra crushed ap12 and pumice to get a good grading of aggragates.
Black pigment.
Hand mixed first then
Mixed in 10kg batches in buckets attached to my rotary tumbler. This was one of the best decisions as mixing 10% is hard work. Orbital sander for vibration.

[charlieslasher]

Form work takes most of the time. Also the quality of the mould and the release agent will save alot of time later in rework and paint. This is the x z axis. Removing the part from the mould was difficult.

[shedbob]

Looks interesting so far.

220kg sounds like a lot,how big will the machine be?

More details on what your building would be of interest too.

Is that a C frame mill column in pictures?

Thanks for posting up the details and progress so far.

Good luck with the project.

[charlieslasher]

[QUOTE=shedbob;1891064]Looks interesting so far.

220kg sounds like a lot,how big will the machine be?

320kg for now. There is still a steel base to cast around that will take an extra 500kg but that will be done later. The machine was steel bolted with epoxy joints and has been running for a few years amd works very good for most aluminium jobs. But I wanted to upgrade the linear rails and increase the z travel.

More details on what your building would be of inteest too.

Is that a C frame mill column in pictures? Its a fixed gantry frame. 600x500y400z

Thanks for posting up the details and progress so far.

Good luck with the project.[
Thanks hopefully some usefull info will come from it./QUOTE]

[charlieslasher]

Major castings all done. After a week of post cure and some tricky demoulding alignment can begin. While still very light weight in machine terms its getting harder to move and position.

[charlieslasher]

Alignment begins with a cast epoxy surface. All moulds were poured on this and gave +-0.08mm surface to start grinding.




Allignment tools guitar string mic and mutimeter good to maybee 0.0025mm with the limit being the mic increments. Resistance spikes as contact between wire and mic begin to seperate. Very repeatable. Over short distace wire sag is small. Still need to calculate how small. Time consuming. Excel spread sheet gives a graphical representation of the linear rail error to help with grinding.



Demoulding was very tricky. I wasted a day trying to get the fist cast part off. Used 4ton jacks, bars and big hammers and nothing would move it. Then I remembered I accidentally designed an air pocket/ram into the mould. 90psi ported to the pocket and it poped off in seconds. About 2ton of evenly distributed force. All parts were cast with air pockets.

[mcphill]

Could you PLEASE elaborate on your guitar string and mic alignment process? That looks very cheap, accurate, and interesting! Have not heard of anyone doing that before...

[charlieslasher]

Could you PLEASE elaborate on your guitar string and mic alignment process? That looks very cheap, accurate, and interesting! Have not heard of anyone doing that before...

Wire is tensioned between ends. In the vertical plane its very strait. In the horizontal plane there is some sag that can be calculated. The longer the axis the more sag
At 0.89m the sag is less than 0.01mm. Set up the linear rail map the errors with the mic. Mic is insulated from the wire/machine. And loss of contact shows up on the mutimeter. I used a machinists level to check/controll roll this is one part thats hard to maintain but can be measured with the wire by of setting the wire from the axis being aligned.

Its quite common with gas turbines and there is some articles from Large hadron collider research that use it over 100m lengths. The hard part is getting the data.

[Eddard_Stark]

My alignment setup on a 3600mm rail was almost identical; one question: I was only using it for a vertical plane as I couldn't find data for wire sag that I was confident in; I was able to find recommended tensioning tables for different size wires. Could you elaborate on how you calculated wire sag?

[charlieslasher]

My alignment setup on a 3600mm rail was almost identical; one question: I was only using it for a vertical plane as I couldn't find data for wire sag that I was confident in; I was able to find recommended tensioning tables for different size wires. Could you elaborate on how you calculated wire sag?

If you can orientate the axis is the vertical plane for measuring then you can get around sag calculations. Furthermore with one axis known to be strait in the vertical plane you can then measure the real sag in the horizontial plane and update the sag model. String tension is the only variable that will change in diffrent setups.

Sag = (W.g/8t)*L^2

W=Wire mass/unit length (17guage guitar string is 0.0114kg/m not measured)
g=9.81 (gravity)
t= tension in wire measured using strings pitch with guitar tuner. Or mass hanging on wire I used a ratchet system.
L is position of measuremnt.
Feed all into a spread sheet and adjust the measured results. Tension of the wire has a large effect on the sag.

Also important measuring from bearing cars averages out errors in the linear way. Good and bad. Good for the final product bad for getting data for alignment.

[Eddard_Stark]

Thx.

[shedbob]

Looking good so far like a serious machine tool.I like it.

I recall the original build now from seeing the pictures.Good to hear it worked well.I am considering epoxy and bolting myself.

The air pockets sound like an effective de-moulding method.Can you show some pictures or describe.

You mention grinding.Is that for rail mounting or are you shimming them? (judging from the pictures)

Thanks.

[charlieslasher]

Grinding shimming where needed and epoxy bonding rails for 100% contact. It takes a week per axis or more depending on how accurate you want to get. A surface plate would help. I have just finished building a laser "bcam" laser webcam tracking system based on Opensource instruments Opensourceinstruments.com design and lwdaq system. Still determining if this will give accurate data but it looks good so far. They have published alot of usefull designs software and data. Making an accurate laser alignment system from parts lying around is quite easy. The software is very good.

Images:

First check with bcam laser alignment.

Z axis with air pocket recessed for release casting from mould.

[charlieslasher]

10k images later I think I have accurate data on all the axis. Bcam is good for measuring roll angle only.

I converted the lasers back to a laser autocollimator with ccd tracking using the bcam software and leds for bcam roll data. Reflectors (HDD) were used to map the axis like standard autocollimators do.

Accuracy and repeatability seem good. Mutiple passes of an axis have tight data distribution. But with no refrence i have no way to calibrate it apart from using a 90arc sec machinists level.

Once a axis is maped grinding and scraping highspots and then remap and repeat.

[shedbob]

So what do you think your system is capable of resolving 1/2 a thou of an inch or better ? Does it allows you to get the surface genuinely coplanar ? I notice it indicates tilt in the graphic.

A cheap Chinese granite surface plate is an excellent investment.I got a class b 12x18x3 inch here in New Zealand delivered for less than $120NZ.Class b is any two points no more than 2/10000 difference.

I recall reading about H-BCAM when its been raised before.Just done another quick skim over it.But "quite easy" wasnt a phrase I would apply to it !!! I have seen a simpler hardware only system in Circuit Cellar some years ago.
Great to see you have a successful system.Sounds like some of your available parts might differ from the average available workshop cast offs though.

[charlieslasher]

The max accuracy of laser ccd tracking im going to call it laserbcam is very good 0.005mm/600mm. Roll measurement with ledbcam +-5arcsec a little better than my starret machinest level. Im working for +-0.005mm.

A refrence straitedge would be usefull for spotting. Right now I colour the way with yellow oil paint and the rail with black die and map the axis assembled. Then scrape the spots from the axis map and repeat. Its very slow job few hundred hours and 40-50 iterations per rail.

A cast iron strait edge would speed it up. Scraping eg is easy however.

But for big routers the laser would be ideal.

[charlieslasher]

Table cast 100mm deep with 7075 plate for top surface. Wired vfd and control in new cabnet. Alignment of z axis to y axis and table to fit.

[charlieslasher]

First cuts have happened. T-sloted table being machined. Waycovers have been difficult to fabricate.

[CanonGuy]

Looks awesome! How has it been working? Any videos of it cutting?

[charlieslasher]

Goes realy good. Eg is perfect material for high speed machining. I get some video. Upgrading spindle to 10hp bt30 currently.